The electrical components of modern computing systems generate significant amounts of heat. Overheating of the components may compromise their performance and/or cause damage thereto. Computing systems therefore typically employ systems for cooling their constituent electrical components.
Cooling systems include and are not limited to fans, heat sinks, and recirculated coolant. Recirculated coolant may be enclosed in a structure which is placed in contact with electrical component to be cooled. Some conventional systems immerse the component within the coolant. For example, a chassis may be sealed and filled with recirculating coolant fluid to cool the components therein, or one or more motherboards may be immersed in a vertically-standing container of coolant.
Immersion-based cooling is suitable for cooling electrical components which are substantially impervious to the utilized coolant. These components include most integrated circuits, discrete components (i.e., capacitors, resistors, etc.) and solid-state hard drives. A conventional hard disk drive, in contrast, regulates internal pressure using a hole which exposes the inside of the drive to ambient air. Such a drive is not compatible with immersion-based cooling because coolant may enter through the hole (or other opening) and thereby render the drive unusable.
Integrated computer servers (including, e.g., a circuit board, memory and data storage) are designed in view of a desired balance of cost and performance. Some server designs favor the cost/performance profile of disk-based drives as compared to solid-state drives. However, disk-based drives are not usable with these server configurations if the servers are to be immersion-cooled. Such immersion-cooled servers require solid-state drives, which are typically more costly than solid-state drives. Systems are desired which facilitate immersion cooling of hard disk drive-based servers.